Crypto Research Day: Crypto x AI Fireside Chat

AI x Crypto Fireside Chat with Emad of Stability AI / Schelling; Alex of Near AI, and Anna of Vana

Crypto Research Day was hosted by ​ Superscrypt ,  Ethereal Ventures  and  ANAGRAM in Singapore on 19th September 2024. As part of the program we had Anna Kazlauskas founder of Vana moderate a fireside chat with Emad Mostaque, the founder of Schelling.ai & Stability and Alex Skidanov, the co-founder of Near AI. Takeaways below

TLDR;

• Just as we can use distributed ledgers to make AI more efficient and aligned, we can use AI to unlock the true promise of Web3, which is self-sovereignty, agency, and the power to make it incorruptible.

• People's data is valuable, though many don't realize it. Most carelessly sign away their data rights, while companies like Facebook generate substantial revenue by selling this information to advertisers. Individuals should have the opportunity to monetize their own data, and there will come a time when they recognize its true worth. Blockchains are poised to play a crucial role in orchestrating this shift.

• Decentralised inference is largely solved already, however it still needs to just gain market adoption. This is at the forefront of Web3 x AI today according to Near. Decentralised compute is not there yet, still in research phase. Once this is figured one, will be a big unlock Web3 and there will be a drastic change in the demand supply here.

Why Web3 x AI?

• Just as we can use distributed ledgers to make AI more efficient and aligned, we can use AI to unlock the true promise of Web3, which is self-sovereignty, agency, and the power to make it incorruptible.

• Open Source vs. Proprietary Models - Proprietary AI is advancing far more rapidly and most major language models and AI systems are closed-source and inaccessible. Only Facebook is making significant efforts to open-source their models. There's a pressing need to develop user-owned AI solutions and this is where blockchains come in. They offer valuable tools for coordination, resilience, and validation in AI systems.

• Open-source AI models are crucial, especially for sectors like healthcare and education, where transparency and accountability are key.

Monetization of Data in a Web3 Ecosystem

• People's data is valuable, though many don't realize it. Most carelessly sign away their data rights, while companies like Facebook generate substantial revenue by selling this information to advertisers. Individuals should have the opportunity to monetize their own data, and there will come a time when they recognize its true worth. Blockchains are poised to play a crucial role in orchestrating this shift.

• Blockchain can enable individuals to monetize their own data, shifting power from centralized platforms to users.

• Data ownership will drive the adoption of Web3, allowing people to participate directly in the digital economy.

• Agent marketplaces and decentralized platforms provide micro-payments and value exchange, enhancing the AI and crypto synergy.

Overcoming Technical Barriers to Decentralized AI

• Decentralized inference is largely solved already, however it still needs to just gain market adoption. This is at the forefront of Web3 x AI today according to Near. Decentralized compute is not there yet, still in research phase. Once this is figured one, will be a big unlock Web3 and there will be a drastic change in the demand supply here.

• Challenges remain in training large AI models on distributed networks, but ongoing research is addressing this.

• Decentralized infrastructure, such as distributed GPUs, is key to scaling AI without reliance on big tech companies.

• The future lies in building high-quality, open-source AI models that can compete with proprietary ones while remaining transparent and accessible.